Welcome

Hi, welcome to my website. On this site there’s an analysis of the class corpus of Computational Musicology 2025.

For my own songs, I found both on a YouTube channel called BORNEO. I’m an amateur DJ, and my usual set starts off with rally house (reinout-w-1) and gradually flows into trance (reinout-w-2). I was looking for royalty hard house at first, but couldn’t find so. Then I started looking for royalty free trance, and BORNEO showed up. On his page, I found both songs.

Chromagram

Cepstogram

Chroma-based self-similarity

Timbre-based self-similarity

The more arousing, the more engaging, the more dancable

<<<<<<< HEAD
=======
>>>>>>> 02828c9 (homework week 9 part 2)

On the left is the graph. When it comes to my tracks, we can see that they are very high in both engagingness and arousal compared to the rest of the class. For example, ‘reinout-w-1’ is number 4 in engagingness, and ‘reinout-w-2’ is number 3 in arousal. They are also both high in danceability.

Essentia characterised my tracks correctly. Both tracks have a lot of energy, which correlates with the variables above being high.

When it comes to the rest of the corpus, my songs have more energy. This is also what I concluded when I listened to the songs of the corpus itself. My tracks are not similar to most of the tracks from the corpus, meaning Essentia did a good job identifying.

What we found

As can be seen in the graph, the engagingness increases when the arousal increases. This means that the engagingness and the arousal are positively correlated. Furthermore, as we can see the engagingness and arousal increase, the color of the data becomes more and more towards yellow, starting from dark blue. This means that the danceabiltiy also increases with the engagingness and the arousal. Conclusion: all three variables are positively correlated.